The output of a fine tuned
gpt-3.5-turbo-1106 model is generating the user turn with the assistant turn before stoping:
ChatCompletionMessage( content='Great. May I have your full name, please?\n[[input_text(usersName)]]', role='assistant', function_call=None, tool_calls=None )
My data is incorrectly formatted. Specifically:
- I did not include a content attribute and value in the assistant messages where a function is called.
- I used the old functions schema of defining functions in the training data and not the tools schema.
Documentation about fine tuning inputs:
- API docs for creating a chat completion. The functions attribute is deprecated.
- Cookbook for functions calling uses the functions attribute with the
- The function calling docs use the functions attribute with a straight
- Not including a
contentattribute with a string value creates an error. Likewise function calls in general create errors.
The docs do not say anywhere definitively what the inputs must look like for a specific model.
I will train a model with the new tools schema but still without the content attribute and see if the issue is resolved.